F1Tenth Racing Reactive Obstacle Avoidance

This project implements a complete Gap Follow algorithm for the F1Tenth autonomous racing simulation using ROS 2 Humble and f1tenth_gym_ros. The objective is to navigate safely through static obstacles using LiDAR scans and reactive control.

---

Overview

Gap Follow is a reactive method where the racer:

• Reads the LiDAR scan
• Identifies all obstacles in its field of view
• Inflates obstacles (safety bubble)
• Detects gaps between obstacles
• Selects the largest & safest gap
• Steers toward the gap center
• Sets speed based on forward clearance & steering

This repository contains:

• Updated simulation using Spielberg obstacle map
• A fully implemented gap_follow.py node
• gap_follow.launch.py that launches f1tenth simulation + controller
• Tuned speed + steering controllers for smooth behavior

---

Gap Follow Algorithm – Implementation Summary

Your implementation (in gap_follow.py) includes:

1. LiDAR Preprocessing

• Subscribes to /scan
• Converts scan into usable ranges
• Clips invalid/inf values
• Restricts view to ±120° FOV

2. Obstacle Inflation

Every detected obstacle creates a “bubble” to avoid clipping corners:

bubble_radius = f(distance_to_obstacle)
inflation_size = min(max(obstacle_inflated, angular_bubble), 60)

All ranges inside the bubble are set to zero (blocked).

3. Gap Detection

A continuous set of non-zero LiDAR points = a gap For each gap, compute:

• Average distance
• Maximum distance
• Gap width
• Center index
• Gap score (weighted by distance + center alignment)

gap_score = 0.4 * gap_max + 0.6 * gap_avg
gap_score *= center_alignment_factor

The highest-scoring gap is selected.

4. Steering Command

Heading direction = angle of center of chosen gap Steering is smoothed:

smoothed = α * desired + (1-α) * previous

Speed-dependent steering reduction prevents over-steering.

5. Speed Controller

Your proportional controller:

• Slows down when gap clearance is low
• Speeds up in open regions
• Scales speed based on steering angle (cosine factor)

speed = Kp * clearance * steering_factor
speed clipped between (min_speed, max_speed)

---

Running the Simulation

1. Build the Workspace

colcon build
source install/setup.bash

2. Launch the Gap Follow System

ros2 launch gap_follow gap_follow.launch.py

This will:

• Start the f1tenth_gym_bridge
• Load the Spielberg obstacle map
• Launch the Gap Follow controller
• Open RViz2 with LiDAR + vehicle visualization

---

Simulation Results

Obstacle-Free Map — Gap Follow

The racer navigates smoothly with high speed.

Obstacle Free Map Gap Follow

---

Obstacle Map — Gap Follow

Obstacle inflation and adaptive gap scoring make the racer safely avoid blocked regions.

Obstacle Map Gap Follow

---

Maps Used

Obstacle Free Map	Obstacle Map

---

Notes

• Uses AckermannDriveStamped for control
• Runs at 10 Hz publish rate
• ROS 2 Humble required
• Must install:

sudo apt install ros-humble-ackermann-msgs

Resources

• F1Tenth Follow-The-Gap - Video
• f1tenth gym ros
• AckermannDriveStamped